The article “Algebraic Reconstruction Techniques Can Be Made Computationally Efficient” by G. T. Herman, IEEE Transactions on Medical Imaging, Vol. 12, No. 3, September 1993 discloses an iterative algebraic reconstruction technique (ART). Starting from an initial image, in each iteration step a) simulated projection data are generated by simulating a forward projection through a current image of an object of interest b), a difference between the simulated projection data and measured projection data is determined, and c) this difference is backprojected for updating the current image of the object of interest.
This reconstruction technique is sensitive to changes in the measured projection data, which can be introduced during a preprocessing step which includes, for example, correction for tube current fluctuations or air calibration. The sensitivity to these changes in the measured projection data may lead to shading artifacts in the reconstructed image and reduces therefore the quality of the reconstructed image.